NightOwl888 commented on issue #403:
URL: https://github.com/apache/lucenenet/issues/403#issuecomment-765044601
@rclabo is partially correct. This requires a custom codec configuration and
the implementation he provided is what I would recommend. However, when
*reading* an index the codec must be properly registered for Lucene.NET to
instantiate it.
When *writing* the index, the codec can be specified as:
```c#
IndexWriterConfig indexConfig = new
IndexWriterConfig(LuceneVersion.LUCENE_48, standardAnalyzer);
indexConfig.Codec = new Lucene46HighCompressionCodec();
```
Or when properly registered (see registration steps below):
```c#
IndexWriterConfig indexConfig = new
IndexWriterConfig(LuceneVersion.LUCENE_48, standardAnalyzer);
indexConfig.Codec = Codec.ForName("Lucene46HighCompression");
```
## Codec Registration
When *reading*, the codec is instantiated based on its text name (in this
case `"Lucene46HighCompression"`) which is stored in the index file. Codecs are
supplied to Lucene.NET using an abstract factory pattern (which fits nicely
with dependency injection if you are using it).
In your application's startup path you will need to add the following line
to register your custom codec with high compression:
```c#
Codec.SetCodecFactory(new DefaultCodecFactory {
CustomCodecTypes = new Type[] { typeof(Lucene46HighCompressionCodec) }
});
```
> Alternatively, see the [codec
documentation](https://lucenenet.apache.org/docs/4.8.0-beta00013/api/core/Lucene.Net.Codecs.html#using-microsoftextensionsdependencyinjection-to-register-a-custom-codec)
for examples of using `Microsoft.Extensions.DependencyInjection` to register
codecs. Registering with other dependency injection containers is similar.
> **NOTE:** Codecs are registered as singleton lifestyle. That is, once
instantiated the same instance with that name is always used for all read/write
operations.
This tells Lucene.NET what type to load. The `CodecNameAttribute` that
decorates the custom codec (which is optional) specifies the name of the codec.
> **IMPORTANT:** since high compression is binary incompatible with normal
compression, you should not reuse the same name as `"Lucene46"` for your custom
codec name. Reusing the same name is allowed and the last registration wins,
but will only work if the codec has the exact same binary format.
----------------------------------------------------------------
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
For queries about this service, please contact Infrastructure at:
[email protected]
